Vector CLI

Vector CLI is an advanced command-line interface developed by Vector Technologies, designed to empower users with seamless model deployment and management capabilities directly from their terminal. With Vector CLI, users can efficiently interact with Vector’s robust platform to handle tasks such as model deployment, configuration, version control, and real-time monitoring, all within a streamlined command-line experience. This tool is ideal for developers, data scientists, and engineers who require a high level of control and integration for managing machine learning models in a fast, intuitive way.

Installation

To install the CLI, use npm:

npm i vector-hub-cli

Table of Contents

• Introduction
• Quick Start
• Commands Overview
• Detailed Commands
• Configuration
• Error Handling
• Examples
• Troubleshooting
• Contributing

Quick Start

Once installed, try running your first command to see a list of all available options:

vector --help

Commands Overview

• A list of available commands, with a brief description of each. This can serve as a quick reference.

• deploy - Deploys a model at the specified path.

• list - Lists all deployed models in the current environment.

• delete - Deletes a model based on its unique ID.

• status - Displays the deployment status of a specific model.

• logs - Fetches and displays the logs for a model.

Detailed Commands

Deploy

Deploy a model to Vector's platform.

Usage:

vector deploy <modelPath> [--name <name>] [--version <version>]

• name : Optional custom name for the deployment.
• version : Specify a version number if deploying multiple versions

List

Lists all models currently deployed.

Usage:

vector list [--verbose]

Configuration

Details on any configuration files or environment variables required for custom setups.

Vector CLI requires a configuration file, vector.config.json, to manage deployments and other settings.

Sample vector.config.json file:

{
    "api_key": "your_api_key_here",
    "project_id": "your_project_id_here",
    "default_env": "production"
}

Error Handling

• Information on common errors, possible causes, and solutions.

If you encounter issues, refer to the error codes below:

• Error Code 101: "Model path not found" - Verify the specified model path exists.
• Error Code 403: "Unauthorized Access" - Ensure your API key is valid and has the necessary permissions.

Examples

Deploying a New Model Version

vector deploy ./models/updated_model --version "2.0" --name "UpdatedSentimentModel"

Checking Model Status

vector status 12345  # where 12345 is the model ID

Troubleshooting

A guide for solving common installation, configuration, and usage issues.

CLI Not Recognized

• Ensure vector is globally accessible by verifying its installation location or reinstalling with npm install -g.

Permission Issues

• If permission errors occur, try running with sudo, or ensure your vector.config.json has the correct API keys.

Contributing

Contributions are welcome! Please see the CONTRIBUTING.md file for guidelines.